the need for integrated life cycle sustainability analysis ......chapter 7 the need for integrated...

Chapter 7

The Need for Integrated Life Cycle SustainabilityAnalysis of Biofuel Supply Chains

Anthony Halog and Nana Awuah Bortsie-Aryee

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/52700

1. Introduction

Climate change has been widely investigated by the scientific community, and its potentialimpacts are expected to affect the world’s economy, ecosystem services, and societal struc‐tures within a few decades. To reduce the undesirable consequences of climate change,adaptation and mitigation technologies and policies have to be implemented. Analyzingtechnological advances to address sustainable development, through integrated systems lev‐el methods and approaches, is needed to predict future vulnerability to climate change andcontinued ecosystem deterioration [1-4]. The incorporation of sustainability notion into allsub-systems of our global society has come into full swing and must be continued to be pur‐sued by any entity, both in private as well as public sectors. However, these challenges re‐quire integrative and transdisciplinary computational tools and methods to aid inembedding sustainability goals into corporate and government policy decision makingprocesses [5-8]. United States Environmental Protection Agency (USEPA) has been reshap‐ing its strategies and programs in conjunction with incorporating the triple dimensions ofsustainability [9].

In the U.S. bioenergy development, recent innovations in biotechnology, genomics and com‐plexity science have contributed to the renewed interest in converting (ligno) cellulosic bio‐mass to valuable fuels and other bioproducts [10]. The U.S. Department of Agriculture(USDA) and Department of Energy (USDOE) are actively supporting projects to make bio‐fuels and bioproducts economically, socially and environmentally sustainable and viable.U.S. Bioenergy Research Centers work on accelerating genomics-based systems biology re‐search to achieve the transformational breakthroughs in basic science needed for the devel‐opment of cost-effective technologies to make production of next-generation biofuels fromlignocellulose, or plant fiber, commercially viable on a national scale [11]. Figure 1 shows

© 2013 Halog and Bortsie-Aryee; licensee InTech. This is an open access article distributed under the terms ofthe Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permitsunrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

the many potential production pathways to biomass hydrocarbons. Features of these alter‐native pathways include diversity in feedstocks, fuel composition, and byproducts. Integrat‐ed decision-making tools are urgently needed to support choices among these alternatives.Developing these tools effectively requires a life-cycle and dynamic perspective. Life CycleAssessment (LCA) follows internationally accepted methods (ISO 14040 and ISO 14044) andpractices to evaluate requirements and impacts of technologies, processes, and products soas to determine their propensity to consume resources and generate pollution.

Figure 1. Pathways to biomass hydrocarbons. Reproduced with permission from [10]; published by Science, 2009

Mckone et al. [1] have identified 7 grand challenges that we must address to enable life cycleassessment (LCA) to effectively evaluate the environmental “footprint” of biofuel alterna‐tives and to support the evolving bioeconomy. According to their work, the grand challeng‐es for applying life cycle assessment to biofuels are:

• Understanding feedstock growers, options, and land use;

• Predicting biofuel production technologies and practices;

• Characterizing tailpipe emissions and their health consequences;

• Incorporating spatial heterogeneity in inventories and assessments;

• Temporal accounting in impact assessments;

• Assessing transitions and end states; and

• Dealing with uncertainty and variability.

The above challenges have already been addressed and analyzed in disaggregated andpiecemeal fashion in several papers in various disciplines. What we really need is to integratethe disparate methodologies systematically and computationally to obtain comprehensive and robust

Biofuels - Economy, Environment and Sustainability198

results (i.e. indicators and metrics) which can support decisions and policies in public and privatesectors. Thus, our one grand challenge is to synthesize our current results and infer relevantand critical information either to support or not bioeconomy development. Integrated-ori‐ented decision-making frameworks and tools with the support of information and commu‐nication technologies (ICT) and cyberinfrastructure are needed to support choices amongthe competing production pathways and provide information to various stakeholders. Forinstance, systems modeling (such as the use of DOE’s GREET model) is relatively inexpen‐sive to perform prior to physically implementing any experiments in laboratory or pilotscale. Towards developing a decision support system (DSS) tool that accounts for multi-stakeholders’ interests in analyzing system’s sustainability, this paper aims to highlight anddescribe the available tools, models and frameworks which can be used to address the 7challenges identified and explore the possibility of their complementary attributes to resultto an integrated system framework for assessing the sustainability implications of further in‐vestments in biofuels, bioenergy and bioproducts.

Thus, an in-depth analysis is needed to critically understand the re-emergence of biofuels inconjunction with sustainable development. Most of the research projects on cellulosic bio‐fuels have not given due attention to social acceptability and economic feasibility, besidesnot considering the competing interests of various stakeholders [2- 3]. The interactions be‐tween environmental, social and economic impacts of biorefinery development must be ana‐lyzed in a comprehensive and integrated manner to ensure the sustainable development of abio-industry (such as the increasing interest in “drop-in” biofuels). Concerns related to bio‐mass harvest and its impact on soil erosion, nutrient losses, biodiversity losses, land usechanges, water consumption, eutrophication and environmental impacts of auxiliaries in‐puts must be weighted with the benefits of cellulosic biofuels [2, 12]. Several studies (e.g.[11, 13-15]) have been conducted to identify biomass potential, assess technological efficien‐cy and understand the environmental implications of biofuels. However, a complete andcomprehensive evaluation of biofuels supply chain from a holistic and systems perspectivehas been lacking. Several authors [2-3; 6-8] have argued the need for further research forlarge scale deployment of second-generation and third generation biomass crops includingtheir effects on land use, biodiversity and hydrology. Technological researches have focusedon using cheap and easily available feedstock (e.g. woody biomass, harvest residues, agri‐cultural residues) to advance lignocelluloses feedstock bio-refinery [16 - 17], whereas envi‐ronmental concerns of biofuels have focused on carbon dioxide emissions only [18]. Thoughintegrated forest products bio-refinery systems may result to additional revenues by pro‐ducing co-products like biofuels and other biomass based chemicals in addition to the mainproducts [16, 19], uncertainties prevail regarding the capital investments required as well asthe social impacts for large-scale production [2-3].

With the above challenges in implementing bioenergy policies while carefully consideringstakeholders’ interests, dynamic integrated system methodologies are urgently needed toanalyze the sustainability of biofuels supply chain and to better understand the overall im‐pacts of introducing (ligno) cellulosic ethanol [5-7; 11] or even to support the developmentof “drop-in” biofuels.

The Need for Integrated Life Cycle Sustainability Analysis of Biofuel Supply Chainshttp://dx.doi.org/10.5772/52700

199

In the subsequent sections of this paper, we elaborate on how we might support the conductof integrated sustainability analysis and modeling of biofuel supply chains.

2. Status

2.1. Understanding stakeholders and resource requirements

The future feedstocks for biofuels may come from farms, rangelands, or forests. Because oftransportation costs, harvested feedstocks are likely to be stored and processed at small- tointermediate-scale (distributed) facilities. Unlike oil companies and government agenciesthat have a hierarchical structure for decision-making, the first stages of biomass produc‐tion might involve hundreds to thousands of decision-makers (stakeholders). Using life cy‐cle perspective approach to influence policies that would alter the behaviors of thesedistributed decision-makers poses different challenges than when the decision-making au‐thority is more highly concentrated. One expects that feedstock growers utilize land tomaximize profits.

Having a large number of potential feedstocks with different characteristics in a system ofdistributed decision-making presents substantial challenges for current LCA approachesbecause of the vast scope of information needed to address so many alternatives. Multi-criteria decision analysis (MCDA), such as the Analytic Hierarchy Process (AHP), can aidin determining the most critical criteria, variables and indicators to stakeholders, whichcan represent their conflicting interests with respect to economic, environmental, techno‐logical and social dimensions of systems sustainability [4-5, 7-8]. The critical criteria andindicators can be ranked and identified by AHP’s eigenvalues, which are calculated fromstakeholders’ inputs [20].

2.1.1. Multi-criteria Decision Analysis (MCDA) for Stakeholders’ analysis

MCDA is a decision support system that is suitable for addressing complex problems featur‐ing high uncertainty, conflicting objectives, different forms of data and information, multi‐ple interests and perspectives, and including complex evolving bio-physical and socio-economic problem [21-26]. MCDA approaches have long been widely applied to economic,social, and industrial systems. An MCDA in general involves m alternatives (e.g. bioenergysystems) evaluated on n criteria (i.e. sustainability criteria), in which each of j-th criteria C ofi-th alternative A has performance of xij. Each criterion is weighted, and wj is the weight ofcriterion j. The grouped (i.e., stakeholders) decision matrix X can be expressed as shown onFigure 2.

Wang et al. [27] present a review of MCDA methods to aid in sustainable energy decisionmaking. In their review, the corresponding methods in different stages of multi-criteria deci‐sion-making (i.e., criteria selection, weighting, evaluation and final aggregation) are dis‐cussed. The criteria are classified into four major aspects: (i) technical (e.g., efficiency,primary energy ratio, etc.), (ii) economic (e.g., investment cost, net present value, etc.), (iii)


environmental (e.g., CO2 emission, NOx emission), and (iv) social (e.g., social acceptability,job creation, etc.). The weighting methods of the criteria are classified into three categories:(i) subjective weighting (e.g., pair-wise comparison, analytical hierarchy process (AHP),etc.), (ii) objective weighting (entropy method, technique for order preference by similarityto ideal solution, etc.), and (iii) combination method.

Criteria c1 c2 ⋯ cn 1

Weights w1 w2 ⋯ wn 2

3

Alternatives A1 X = x11 x12 ⋯ x1n 4

A2 x21 x22 ⋯ x2n 5

⋮ ⋮ ⋮ ⋱ ⋮ 6

An xm1 xm2 ⋯ xmn 7

Figure 2. Grouped decision matrix of MCDA

Different MCDA approaches have been applied to support different decisions including en‐vironmental and sustainable energy decision making [28-30]. We can use any of MCDAmethods to aid stakeholders’ analysis to find out the “most critical” criteria, indicators andmetrics that represent stakeholders’ interests. A combination of AHP and LCA has beenused for evaluating environmental performance of pulp and paper manufacturing [31]. Hal‐og [20] and Halog et al [29] proposed the use of analytic hierarchy process (AHP), one ofMCDA methods [32-34] in stakeholders’ analysis for identifying the critical criteria, indica‐tors and metrics which represent multi-stakeholder’s interests. This will provide ranking ofdifferent criteria, indicators and metrics which are important holistically. MCDA opensgreat applicability to support sustainability assessment of existing and emerging multi-at‐tribute systems. For instance, AHP allows stakeholders to weigh different criteria, indica‐tors, and metrics by calculating Eigen values. In biofuels system, where energy efficiency,investment cost, GHG emissions, land use change and social impacts are the most commoncriteria, MCDA is certainly applicable. Through MCDA, we can focus first on the criticalones that account stakeholder’ inputs.

However, the existing life cycle thinking and MCDA methods are considered steady-statemethods whereby they provide snapshots of hotspots based on historical data. They do notprovide projections or trends in the future. They do not take into account the interactions ofdifferent metrics, outputs and parameters over time. To make the results more useful for de‐cision and policy makers, we need to model the dynamic interrelationships of these varia‐bles over time. Additionally, we can explore the use of Geographic Information Systems(GIS) to assist spatial analysis when needed.

2.2. Biofuel production technologies and practices

Much of the variability among LCA results for biofuels arises from lack of knowledge abouthow biomass production operations and fuel production from biomass will evolve. Many al‐


201

ternatives exist both for production processes and for final products. An important challengeis to understand the energy, biomass, pollutant, and product mass balances of production fa‐cilities: To what extent will they be self-sufficient or even net producers of electricity? Will thefacilities deliver a single product (fuel) or have multiple product streams (fuel, food, electrici‐ty, chemical commodities)? What are their waste products, air emissions, and water de‐mands? A related challenge is accurately predicting scales of future biofuel production. Forbiofuels, the feedstocks are more dispersed and less dense than petroleum, which will inducebiorefineries to be smaller than petroleum refineries. Ultimately, the scale of biorefining willdepend on feedstock and production process choices, technological efficiency in convertingfeedstock to fuel, productivity of local land for feedstock production, and costs associatedwith feedstock production and transport, and biorefinery construction and operation. Much isunknown about this system at large scale and will remain uncertain until the system is creat‐ed. Larger biorefineries may economize on refining-related impacts, but will increase trans‐port-related impacts. Biorefinery scale has important ramifications for life-cycle impactsincluding the nature and the location of impacts. We can use the scenario analysis capabilityof the methods of agent based modeling (ABM) and system dynamics (SD) to predict poten‐tial performance of various biofuel production technologies and practices.

2.2.1. Agent Based Modeling (ABM)

Agent based model is a computer representation of the considered system (e.g. biofuels sup‐ply chain) that is comprised of multiple, interacting actors (i.e., stakeholders) [35]. ABM sys‐tems possess two distinct properties: (1) the system is composed of interacting agents; and(2) the system exhibits emergent properties, that is, properties arising from the interaction ofthe agents/stakeholders that cannot be deduced simply by aggregating the properties of theagents [36- 37]. ABM can be used to model the interactions of agents or sub-systems in bio‐fuels supply chain using the metrics, variables and indicators as performance measures. Fig‐ure 3 provides schematic illustration of an agent-based system: each of the four circlesrepresents a sub-system of agents (e.g., companies/entities) denoted by small dots and thewhole arrows show how agents and sub-system of agents are interacting with each other.Interacting agents and sub-systems, though driven by only a small set of rules which governtheir behavior, account for complex system behavior whose emergent dynamic propertiescannot be explained by analyzing its component parts [35].

ABM aims to look at global consequences of individual or local interactions in a given geo‐graphical (e.g. regional) area. Parker et al [38] have categorized existing literatures on agent-based land use models into five categories—(i) policy analysis and planning; (ii)participatory modeling; (iii) explaining spatial patterns of land use or settlement; (iv) testingsocial science concepts; and (v) explaining land use functions. Fox et al [39] argue that opti‐mization of supply chain performance is only possible when the impacts of decisions madeby one agent onto another agents are understood. A systems model that captures all impor‐tant interactions among different units of a supply chain would contribute to effective deci‐sion making. Julka et al [40] use petroleum refinery integrated supply chain modeler andsimulator to mimic a crude refinery’s supply chain to develop procurement strategies. Thus,


ABM can be used to look at the global consequences in developing biofuels supply chain/network considering the individual interactions of stakeholders across all life cycle stages.

Figure 3. Illustration of an agent-based system

2.2.2. System Dynamics (SD)

SD is a well-established systems perspective/complexity science method which is originallydeveloped by Jay Forrester at MIT [41- 42]. This has been applied in different corporate, in‐dustrial and government decisions worldwide which have the intention of modeling andunderstanding the interrelationships (i.e., feedbacks) of variables, indicators and metricsover time. This has been useful in modeling the interrelationships between or among sub-systems which are linked by variables and aids to see how their interlinkages will producespecific overall system behavior. Before using appropriate modeling software package, it isimportant to draw causal loop diagrams. A causal loop diagram is a visual representation ofthe feedback loops in a system whereby the stocks and flows (i.e., involving different varia‐bles, parameters, indicators) are connected by either positive and negative loops. A stock(e.g., biomass, GHG, revenue, unemployment) is the term for any entity in the system thataccumulates or depletes over time. A flow is the rate of change in a stock. A flow changesthe rate of accumulation of the stock. The real power of system dynamics is utilized throughsimulation and in showing the inter-linkages between micro-, meso-, and macro-systems.SD involves computational modeling for framing, understanding, and discussing complexissues and problems [43-45]. It is recognized that the structure of any emerging system—themany circular, interlocking, sometimes time-delayed relationships among its components—is often just as important in determining its behavior as the individual components them‐selves. There are often properties of the whole which cannot be found among the properties-of-the-elements. The feedback loops as well as the use of stocks and flows can represent and


203

model the critical sustainability variables, indicators, and metrics to describe how seeminglysimple sub-systems display baffling nonlinearity for the whole system [46]. The modelingcan be developed by sub-dividing the whole system into sub-models but we need to remem‐ber that they are interconnected by variable, parameter or a metric [4, 47]. Through SD wecan create a prototype dynamic system model for the system being considered.

In SD, a system is modeled mathematically in a nonlinear, first-order differential (or inte‐gral) equation such as:

ddt x(t)= f (x, p) (1)

where x is a vector of levels (stocks or state variables), p is a set of parameters, and f is anonlinear vector-valued function. Simulation of such systems is accomplished by partition‐ing simulated time into discrete intervals of length dt and stepping the system through timeone dt at a time.

SD typically goes further and utilizes simulation to study the behavior of systems and theimpacts of alternative policies [44-46]. Running “What If” simulations or scenarios to test cer‐tain energy and environmental policies on a prototype system model can greatly aid in un‐derstanding how an emerging system potentially evolves over time. Similar to MCDA andABM methods, SD has been applied in a wide range of areas, for example population, eco‐logical and economic systems, which usually interact with each other. SD has been used inthe sustainability assessment of technologies in the Canadian Oil Sands Industry [4, 47] aswell as in bioethanol production in Canada [48]. SD models have recently been developed insome biofuels studies. Riley et al [49] use SD model to describe the U.S. DOE biomass pro‐gram. Bush et al [50] and Sheehan [51] explore the potential market penetration scenarios forfirst generation bio-fuels in the United States. Scheffran and BenDor [52] investigate interac‐tion between economic conditions and land competition between different crops. Franco etal [53] use SD to understand the difficulties in fulfilling government requirements for bio‐fuels blending and to evaluate the effect of different government policies in the productionof ethanol and biodiesel.

Once a system model is validated, the procedural steps can be done iteratively dependingon the scenario defined. This will also provide information about the trends of importantvariables over time which will give us insights and guidance on what decisions and policiesto take. Eventually, this modeling procedure can support the selection and implementationof sustainable biofuel systems.

2.3. Characterizing emissions and their health consequences

LCAs of transportation fuel systems report that the fuel combustion stage makes the largestcontributions to pollutant emissions and associated disease burdens. Credible and reliableimpact estimates for biofuel combustion are needed. Another aspect of LCA’s tailpipe chal‐lenge is the need for accurate emission factors for future fleets that cover a range of fuel al‐ternatives and vehicle technologies. Enormous technological progress has been made in


controlling motor vehicle emissions, and there is strong momentum for continuing progress.In what ways and to what extent will shifts from petroleum-based fuels to biofuels affect thecombustion-phase emissions of air pollutants? One historical approach to answer similarquestions has been to conduct laboratory-based emissions testing. However, this approachis relatively expensive and lacks reliability for characterizing fleet-wide emissions from realdrivers on real roads, and so is unlikely to provide accurate information in a timely manner.An alternative approach, used in LCA tools such as GREET [27] which assumes that vehicleemissions meet federal and state emissions standards regardless of the fuel used, and thatemissions targets are the best estimate of what will happen in future years.

Since the concept of sustainability insinuates temporal and spatial connotations, it is impor‐tant that variables, indicators, metrics and parameters representing stakeholders’ interestsare modeled over time within a geographical location, which can be done using dynamicsystem modeling.

2.4. Incorporating spatial heterogeneity in integrated assessment

A key challenge for integrated sustainability assessment of biofuels is to rationally select ap‐propriate spatial scales for different impact categories without adding unnecessary complex‐ity and data management challenges. Though methods such as LCA can address netchanges across large geographical areas, it must also address how the impacts will be expe‐rienced at local or regional scales. Accurate assessments must not only capture spatial varia‐tion at appropriate scales (from global to farm-level), but also provide a process to aggregatespatial variability into impact metrics that can be applied at all geographical scales. We canuse geographic information system (GIS) to support regionalized LCA. GIS has a powerfulanalytical ability to assess data spatially. It stores different layers of information. GIS is atool that links location and attribute information to enable a person to visualize patterns, re‐lationships, and trends for different parameters of a system. Previous literature and organi‐zational reports have suggested using GIS as a tool to link the aspatial data with spatial data[54-56]. This offers at least two important contributions to the modeling of land use changesimpacts on biodiversity and ecosystem goods and services caused by biofuels production. Itpermits analysis of spatially explicit datasets of land use and land cover, which can be usedin the assessment of the areas affected by increased land occupation. It also offers an under‐standing of land use dynamics. The insight can aid in the predictions of land use changescaused by increased demand for biofuels [57].

2.5. Temporal accounting in integrated sustainability assessment

Similar to spatial resolution when conducting integrated sustainability analysis, selectingappropriate time scales poses challenges for biofuels system modeling efforts. Time alloca‐tions are important for comparing impacts, yet the time distribution of impacts is rarelymade clear in studies such as LCAs. Many factors in systems analysis vary significantly intime. Therefore, time-based assumptions must be clearly noted and evaluated. Among the“moving targets” are population distributions, technology options, regulatory requirements,and the degree of biofuel penetration in the overall energy mix. Moreover, the inputs one


205

uses in integrated sustainability assessments to characterize biomass and fuel productiontechnologies as well as transportation infrastructure must capture how these systems areevolving. Different natural time scales associated with different impacts pose challenges foreffective comparisons among climate-change, human health, and water use consequences.The impacts of GHG emissions are distributed over decades using integrated assessmentmodels and are commonly discounted. Assumptions about the rate of discounting influencejudgments about the relative importance of current year versus future year emissions. Simi‐larly, impacts on water resources and soil can play out over decades. We can use dynamicsystem modeling and simulation.

2.6. Assessing transitions and end states in sustainability analysis

Both advocates and critics of biofuels often focus on a restricted set of scenarios that appearto reinforce their a priori beliefs about how biofuel production and use might function [1].Even accomplished practitioners of LCA tend to focus attention on system end-states (relat‐ed to backcasting), i.e., what biofuel production and use will be likely 20 or 30 years fromnow, when a proposed combination of fuels and vehicles has matured and is thoroughly de‐ployed. This perspective ignores potentially important effects that accrue during the transi‐tion phase; the impacts from building new infrastructure, new vehicles, and integrating anew fuel into a mature and, in many respects, inelastic transportation system. This can beaddressed using dynamic system modeling and simulation or agent based modeling as dis‐cussed above.

To account for transitions, LCA requires much collaboration between economists and sys‐tems engineers to address what happens during the transition phase when large-scalechanges occur in many components of a complex, market driven, technological system. Forexample, one of many key issues is whether fuel changes will affect the performance andlifetime of vehicles or the infrastructure transporting that fuel in ways that significantly in‐crease climate forcing, water, health, and other externalities during transition. Technologyinvestments are needed, and these activities could cause GHG emissions to rise in the near-term as part of a longer-term effort to attain a more carbon-efficient end state. In addressingtransitions, there should be recognition that emerging technologies could profoundlychange the assumptions that underlie biofuel LCAs. This issue makes clear the need to sup‐port life cycle thinking methods for building scenarios from which one should learn, ratherthan as a tool designed to make firm predictions.

2.6.1. Scenario development and analysis for policy planning and making

Forecasting, foresighting and backcasting are approaches used for policy planning and mak‐ing. These scenario development approaches have their benefits and shortfalls. Backcastinginvolves working backwards from a particular desired future end-point or set of goals (i.e.,sustainable society) to the present state, in order to determine the physical feasibility of thatfuture and the policy measures that would be required to reach the state [58-60]. This helpsin analyzing alternative futures response to present situation and deals with problems in adifferent way rather than extrapolating present scenario into the future (forecasting) [61-64].


Backcasting makes it clear that addressing sustainability concerns requires a paradigm shiftfrom business-as-usual attitudes. On the other hand, industries and organizations use fore‐casting technique as a data analysis methodology to develop future scenarios from existinginformation. Forecasting enables decision makers to identify reasonable estimates of variouscurrent activities. Using forecasting approach, managers and decision makers can tweak andcalibrate their operations at the appropriate time in order to maximize benefits. Forecastingassists in preventing losses by taking in all relevant information and making proper judg‐ment decisions [63]. Moreover, foresighting can be distinguished from forecasting. Forecast‐ing is the passive attempt to diagnose or predict future events. Foresighting aims to activelychange or create the future by linking it to the present. Thus, the major difference betweenforesighting and forecasting is that in forecasting the conclusions for today are missing.There are four major applications of foresighting: (1) assessing possible consequences of ac‐tions, (2) anticipating problems before they occur, (3) considering the present implications ofpossible future events; and (4) envisioning desired aspects of future societies. Foresightingwhich is a tool for 'decision-shaping' rather than 'decision-making' offers many benefits in‐cluding: engaging policy-makers and experts in actively planning for the future; identifyingpotential problems early; verifying expectations and examining trends; bringing people to‐gether to create a suitable future; strengthening existing networks,; and educating the publicon urgent future-related issues. It could have a positive impact on sustainable technologypolicy by providing a means for analyzing its broader social and economic implications.

By considering different scenarios (starting from business-as-usual to different plausible sce‐narios in the future (either through foresighting or backcasting), we can generate differentresults for our system performance measures and identify a few critical alternative systemsor scenarios to be strongly considered for developing sustainable decision, policy, technolo‐gy, systems, intervention, etc. Again, this is with the assumption that we can gather goodquality data. We can also perform sensitivity and uncertainty analyses to improve the ro‐bustness of integrated sustainability analysis results.

2.6.2. Uncertainty and variability analysis for robust sustainability assessment

Addressing uncertainty is a major hurdle, not only for biofuels LCA, but for other integratedsystem modeling efforts as well. Many sources of uncertainty and variability, both inherentand epistemic, are encountered in climate-change, human-health, environmental, and eco‐nomic impact assessments. Some of the uncertainties and variabilities cannot be reducedwith current knowledge (i.e., through improvements in data collection or model formula‐tion) because of their spatial and temporal scale and complexity. Effective policies are possi‐ble, but such policies must explicitly take into account uncertainty. Among thosecommenting on how to formally address uncertainty in impact assessments, it has been es‐tablished that there are “levels” of sophistication in addressing uncertainty. In its recom‐mendations for addressing uncertainty in risk assessment, the International Program onChemical Safety proposed four levels, ranging from the use of default assumptions to so‐phisticated probabilistic assessment [1]:


207

• Level 0: Default assumptions; single value of result;

• Level 1: Qualitative but systematic identification and characterization of uncertainties;

• Level 2: Quantitative evaluation of uncertainty making use of bounding values, intervalanalysis, and sensitivity analysis; and

• Level 3: Probabilistic assessments with single or multiple outcome distributions reflectinguncertainty and variability.

Furthermore, Baumgartner [65] argued that assessing environmental and social impacts isassociated with uncertainties caused by applied assessment tools, definition of assessmentobjectives, system boundaries of assessment and data quality. There are various ways todeal with data and model uncertainties when conducting system modeling and simulation[6, 66]. Uncertainties and variabilities come from a large number of variables and parame‐ters considered, assumptions made, and the spatial and temporal variability in parametersor sources [67-69]. The latest LCA, MCDA, and system dynamics software packages includestatistical tools to support uncertainty and sensitivity analyses. Uncertainty analysis aids toshow if the model’s general pattern of behavior is strongly influenced by changes in criticalparameters. For system dynamics, the usual method is to perform a sensitivity analysis ofthe model whereby a collection of simulated experiments is performed [70]. This is done bychoosing parameters, metrics and indicators that are judged to be sensitive, changing theirvalues and then re-running the simulation model. If there is a drastic response in the results,this can show a lack of robustness in the system model. For ABM, the goal is to checkwhether the model addresses the right problem and provides accurate information aboutthe system being modeled [71]. Additionally, Miller [72] proposes to use computer-basedActive Nonlinear Tests (ANTs) that are capable of performing multivariate sensitivity analy‐sis, model breaking and validation, extreme cases, and policy discovery. ANTs search acrosssets of parameter values and are capable of detecting important non-linear relationshipsamong the parameters—relationships that typically go unnoticed using standard techniques[73]. Besides the probabilistic approach in handling uncertainty, fuzzy mathematics, MonteCarlo simulation, etc. can be used to address imprecision [30]. As we confront uncertaintyand variability, we need to separate the “doable” and “knowable” from assumptions thatare conditional components of the integrated and life cycle sustainability assessments. Aninformative system perspective assessment should sort out the data gaps that can be ad‐dressed with modest effort from those that would require a major undertaking. All integrat‐ed assessment efforts require tools, such as sensitivity analysis, variance propagationmethods, and decision/event trees for tracking the impact of data quality and model uncer‐tainty through all components of an assessment. A strong challenge for integrated assess‐ments in addressing uncertainty is to provide and track metrics of data quality with respectto how data were acquired (measurements, assumptions, expert judgment, etc.), to what ex‐tent the data have been validated or corroborated, and how well the data capture technolog‐ical, spatial, and temporal variations. This can be best facilitated by capitalizingcyberinfrastructure such as web-based data mining techniques.


3. Conclusions and discussion

Confronting the seven grand challenges, as argued by McKone et al [1], means recognizingsome issues that have not been well articulated among practitioners of integrated sustaina‐bility assessment. In particular, a good balance must be attained between the needs of tech‐nology momentum and adaptive decision making. Most importantly, we must recognizethat integrated assessment is an adaptive and ongoing process and not just a product. Tech‐nology momentum refers to the difficulties encountered in backing away from fixed costs(financial, institutional, and environmental) that have been sunk into one alternative path‐way. Adaptive decision-making refers to learning by doing, recognizing that commoditycosts and impacts can diminish as a system scales up. For biofuels we need technology mo‐mentum, but we must simultaneously maintain options for adaptive decision-making.These can be handled by the system-based approaches proposed in this paper which are be‐ing currently applied to wood-derived biofuels.

Although life cycle thinking methods can provide insights on options with the lowest im‐pacts, results are often burdened by uncertainty such that they become more informative astechnologies are deployed, making it difficult to apply life cycle thinking methods duringthe early phases of a major technology shift. One example is the important decision thatmust be confronted in a transition to (ligno) cellulosic biofuels; what end-product should betargeted among choices such as alcohols, alkanes, or a specific chemical compound such asdimethyl furan? This type of decision hinges on issues of timing, technical feasibility, andcompetitive advantage.

More collaboration and dialogue between basic scientists and sustainability practitioners isimportant for incorporating system thinking concepts into early phases of technology evalu‐ation. Overall, approaches are needed to create more cross-talking among all members ofthe biofuel enterprise, which could be implemented at web-based level. Ideally, efforts to‐ward developing the science and technology of biofuels will be continuously informed bythose who are expert in integrated sustainability assessment. In this way, the biofuels com‐munity has the best opportunity to attain the overarching sustainability goals they seek. Inintegrated assessment of biofuels, one must recognize that no large-scale industrial productcan be developed in isolation. Natural resource systems such as food, energy, water, andland are all intimately interconnected and thus integrated dynamic system modeling andanalysis is absolutely needed.

To confront the uncertainty associated with sustainability analysis of biofuels combinedwith the irreducibility of many uncertainties, planners and policy makers must considertheir role in managing uncertainty as well as managing impacts. Managing uncertainty re‐quires addressing different aspects of the overall decision making process in the context ofuncertainty. For example, decisions must be made to allocate resources among (i) invest‐ments to collect, store, and manage information; (ii) investments to improve the knowledgebase (i.e., to generate new knowledge); (iii) formalization of the processes used to collect,


209

use, and process information; (iv) formalization of processes to evaluate and communicateuncertainty; and (v) adjustment of the risk assessment process to mitigate the practical im‐pact of the uncertainty on the analysis process. These tasks can be facilitated with the sup‐port of ICT and cyberinfrastructure.

Barriers to integrated sustainability assessments will be similar to environmental LCA (suchas many stakeholders want a final answer), to be “cleared for takeoff, with no call-backs.These stakeholders view these assessments as a final exam; pass it and you are done beingconcerned with impacts and can proceed to technology deployment. This conceptualizationserves to highlight the flaws in our thinking in sustainable technology development andcommercialization. As McKone et al [1] pointed out this shortsighted perspective will fail totake advantage the true power of integrated and life cycle sustainability assessments. At itsbest, integrated sustainability assessment contributes to an ongoing process that organizesboth information and the process of prioritizing information needs. These are opportunitiesfor sustainability practitioners to focus attention and effort on making integrated, systemperspective assessments more useful to decision makers. Integrated and life cycle sustaina‐bility analysis methods can co-evolve with a technology and provide the basis for adaptiveplanning. Decision makers who work in real time and often cannot wait for precise resultsmust recognize that integrated sustainability assessment can provide valuable insights but itis not necessarily a “truth generating machine”. Effective integrated and life cycle sustaina‐bility assessments can guide and inform decisions, but they cannot replace the wisdom, bal‐ance, and responsibility exhibited by effective decision-makers.

Acknowledgement

This work was carried out as part of the seed grant supported by the United States Depart‐ment of Agriculture (USDA) Sustainable Bioenergy Program.

Author details

Anthony Halog1,2* and Nana Awuah Bortsie-Aryee1

*Address all correspondence to: [email protected]

1 School of Forest Resources, University of Maine, Orono, Maine, USA

2 School of Geography, Planning and Environmental Management, University of Queens‐land, Brisbane, Queensland, Australia


References

[1] McKone T, Nazaroff W, Berck P, Auffhammer M, Lipman T, Torn M, Masanet E,Lobscheid A, Santero N, Mishra U. Grand challenges for life-cycle assessment of bio‐fuels. Environmental Science and Technology; 2011.

[2] Rowe RJ, Terry RC, Rickart EA. Environmental change and declining resource availa‐bility for small-mammal communities in the Great Basin. Ecology 2011; 92(6),1366-1375.

[3] van Dam J, Junginger M, Faaij A, Jurgens I, Best G, Fritsche U. Overview of recentdevelopments in sustainable biomass certification. Biomass and Bioenergy 2008;32(8), 749-780

[4] Halog A, Chan A. Developing a Dynamic Systems Model for Sustainable Develop‐ment of the Canadian Oil Sands Industry, International Journal of EnvironmentalTechnology and Management 2008; 8 (1): 3-22.

[5] Halog A, Manik Y. Advancing integrated systems modeling framework for life cyclesustainability assessment. Sustainability 2011; 3 (2):469-499

[6] Halog A. Models for evaluating energy, environmental and sustainability perform‐ance of biofuels value chain. International Journal of Global Energy Issue 2009; 32:87-101.

[7] Buchholz T, Luzadis V, Volk T. Sustainability criteria for bioenergy systems: Resultsform an expert survey. Journal of Cleaner Production 2009; 17: S86-S97.

[8] Buchholz T, Rametsteiner E, Volk T, Luzadis V. Multi criteria analysis for bioenergysystems assessments. Energy Policy 2009; 37:484-495.

[9] National Academy of Sciences. Expanding Biofuel Production: Sustainability and theTransition to Advanced Biofuels: Summary of a Workshop. National Academy Press;2010

[10] Regalbuto J. Cellulosic Biofuels – Got Gasoline? Science 2009; 325: 822-824.

[11] USDA/USDOE. Sustainability of Biofuels: Future Research Opportunities. Reportfrom the October 2008 Workshop, April 2009; 2009.

[12] Davis SC, Anderson-Teixeira KJ, DeLucia EH. Life-cycle analysis and the ecology ofbiofuels. Trends in plant science 2009; 14(3): 140-146.

[13] Farrell AE, Plevin RJ, Turner BT, Jones AD, O’Hare M, Kammen DM. Ethanol cancontribute to energy and environmental goals. Science 2006; 311:506–508.

[14] Plevin RJ, O’Hare M, Jones AD, Torn MS, Gibbs HK. Greenhouse gas emissions frombiofuels’ indirect land use change are uncertain but may be much greater than previ‐ously estimated. Environment, Science and Technology 2010; 44: 8015–8021.


211

[15] Perlack R, Wright L, Turhollow A, Graham R, Stokes B, Erbach D. Biomass for a feed‐stock for bioenergy and bioproducts industry: the technical feasibility of a billion tonannual supply, Oak Ridge National Laboratory, TN (2005 May) 2005;78 p. ORNL/TM-2005/66, DOE/GO-102995, 2135.

[16] Mao H, Genco JM, Yoon SH, van Heiningen A, Pendse H. Technical economic evalu‐ation of a hardwood biorefinery using the "Near-Neutral" hemicellulose pre-extrac‐tion process. Journal of Biobased Materials and Bioenergy 2008; 2(2), 177-185. doi:10.1166/jbmb.2008.309

[17] Uihlein A, Schebek L. Environmental Impacts of a Lignocelluloses Feedstock Biorefi‐nery System: An Assessment. Biomass and Bioenergy 2009; 33: 793-802.

[18] Dickerson K., Rubin J. Maine Bioproducts Business Pathways. Margaret Chase SmithPolicy Centre, http://www.forestbioproducts.umaine.edu/library/Maine-Bioprod‐ucts-Business-Pathways/203/. Forest Bioproducts Research Initiative, School of Eco‐nomics, University of Maine, Orono ME; 2008.

[19] Halog A, Mao H. Assessment of Hemicellulose Extraction Technology for BioethanolProduction in the Emerging Bioeconomy. International Journal of Renewable EnergyTechnology 2011; 2(3): 223 - 239.

[20] Halog A. Sustainable development of bioenergy sector: An integrated methodologi‐cal framework. International Journal of Multicriteria Decision Making 2011. In press.

[21] Pór A, Stahl J, Temesi J. Decision support system for production control: Multiple cri‐teria decision making in practice. Engineering Costs and Production Economics 1990;20: 213-218.

[22] Spengler T, Geldermann J, Hähre S, Sieverdingbeck A, Rentz O. Development of amultiple criteria based decision support system for environmental assessment of re‐cycling measures in the iron and steel making industry. Journal of Cleaner Produc‐tion 1998; 6: 37-52.

[23] Munda G. A NAIADE based approach for sustainability benchmarking. Internation‐al Journal of Environmental and Technology Management 2006; 6:65-78.

[24] Munda G. Social multi-criteria evaluation: methodological foundations and opera‐tional consequences. European Journal of Operations Research 2003; 3: 662-677.

[25] Roy B. The Outranking Approach and the Foundations of ELECTRE Methods; Bana eCosta, C.A., Ed.; Springer-Verlag: Berlin, Germany; 1991: pp155-183.

[26] Brans J, Vincke P. PROMETHEE method for multiple criteria decision-making. Man‐agement Sciences 1985; 31:647-656.

[27] Wang JJ, Jing Y-Y, Zhang CF, Zhao JH. Review on multi-criteria decision analysis aidin sustainable energy decision-making. Renewable and Sustainable Energy Review2009; 13: 2263-2278.


[28] Pohekar S, Ramachandran M. Application of multi-criteria decision making to sus‐tainable energy planning—a review. Renewable and Sustainable Energy Reviews2004; 8: 365-381.

[29] Halog A, Wei B, Sagisaka M, Inaba A. A Multi-attribute assessment of environmen‐tally-sound electric vehicle battery technologies. Journal of Industrial Engineering2004; 1: 40-59.

[30] Halog A. Selection of Sustainable Product Improvement Alternatives. Ph.D. Disserta‐tion. http://digbib.ubka.uni-karlsruhe.de/volltexte/242002 (accessed 16 February2011). University of Karlsruhe, Karlsruhe, Germany, 2002.

[31] Pineda-Henson R, Culaba A, Mendoza G. Evaluating environmental performance ofpulp and paper manufacturing using the analytic hierarchy process and life-cycle as‐sessment. Journal of Industrial Ecology 2002; 6: 15-28.

[32] Saaty TL. The Analytic Hierarchy Process; McGraw Hill Company, New York, NY,USA; 1980.

[33] Saaty TL. How to Structure and Make Choices in Complex Problems. Human Sys‐tems Management. 1982; 3-4: 255-260.

[34] Saaty TL. Theory and Applications of the Analytic Network Process. RWS Publica‐tions, Pittsburgh, PA, USA; 2005.

[35] Holland JH, Miller JH. Artificial adaptive agents in economic theory. American. Eco‐nomic Review 1991; 81: 365-371.

[36] Akkermans, H. Emergent supply networks 2001: System dynamics simulation ofadaptive supply agents. In Proceedings of the 34th Hawaii International Conferenceon System Sciences Hawaii, HI, USA, 3–6 January 2001; p. 11.

[37] Axelrod, R. The Complexity of Cooperation: Agent-Based Models of Competitionand Collaboration; Princeton University Press: Princeton, NJ, USA; 1997.

[38] Parker D, Manson S, Janssen M, Hoffmann M, Deadman P. Multi-agent systems forthe simulation of land-use and land-cover change: A review. Annals of the Associa‐tion of American Geographers 2003; 93: 314-337.

[39] Fox MS, Barbuceanu M, Teigen R. Agent-oriented supply-chain management. Inter‐national Journal of Flexible Manufacturing Systems 2000; 12:165-188.

[40] Julka N, Karimi I, Srinivasan R. Agent-based supply chain management: A refineryapplication. Computers and Chemical. Engineering 2002; 26:1771-1781.

[41] Sterman JD. System dynamics modeling: Tools for learning in a complex world. Cali‐fornia. Management Review 2001;43: 8-25.

[42] Borshchev A, Filippov A. From system dynamics and discrete event to practicalagent based modeling: Reasons, techniques, tools. In Proceedings of the 22nd Inter‐national Conference of the System Dynamics Society, 2004. Oxford, UK, 25–29 July.


213

[43] Radzicki M, Taylor R. Origin of System Dynamics: Jay, W. Forrester and the Historyof System Dynamics. http://www.systemdynamics.org/DL-IntroSysDyn/start.htm(accessed 22 February 2011). U.S. Department of Energy‘s Introduction to System Dy‐namics; 2011.

[44] Forrester JW. Industrial Dynamics. John Wiley & Sons Inc, New York, NY, USA;1961.

[45] Forrester JW. Counterintuitive behavior of social systems. Technology Review1971;73: 52-68.

[46] Saeed K. Slicing a complex problem for systems dynamics modeling. System Dynam‐ics Review 1992; 8: 251-262.

[47] Halog A, Chan A. Toward sustainable production in the Canadian oil sands indus‐try. In Proceedings of the 13th CIRP International Conference on Life Cycle Engineer‐ing, European Union, 31 May–2 June 2006; pp. 131-136.

[48] Chan A, Hoffman R, McInnis B. The role of systems modeling for sustainable devel‐opment policy analysis: The case of bio-ethanol. Ecology and Society 2004; 9, Article6.

[49] Riley C, Wooley R, Sandor D. Implementing systems engineering in the US. Depart‐ment of Energy Office of the Biomass Program. http://www.nrel.gov/docs/fy07osti/41406.pdf (accessed 16 February 2011). In Proceedings of International Conference onSystem of Systems Engineering: SOSE in Service of Energy and Security, San Anto‐nio, TX, USA, 16–18 April 2007; 2007

[50] Bush B, Duffy M, Sandor D, Peterson S. Using system dynamics to model the transi‐tion to biofuels in the United States. In Proceedings of the Third International Confer‐ence on Systems of Systems Engineering Monterey, California, CA, USA, 2–4 June2008; NREL/CP-150-43153. http://www.nrel.gov/docs/fy08osti/43153.pdf (accessed 16February 2011).

[51] Sheehan JJ. Biofuels and the conundrum of sustainability. Current Opinion in Bio‐technology 2009; 20: 318-324.

[52] Scheffran J, BenDor T. Bioenergy and land use: A spatial-agent dynamic model of en‐ergy crop production in Illinois. International Journal of Environmental Pollution2009; 39: 4-27.

[53] Franco C, Ochoa MC, Florez AM. A system dynamics approach to biofuels in Colom‐bia. In Proceedings of the 27th International Conference of the System Dynamics So‐ciety 2009, Albuquerque, NM, USA, 26–30 July 2009. http://www.systemdynamics.org/conferences/ 2009/proceed/papers/P1189.pdf (accessed 16February 2011).

[54] Keam S. McCormick N. Implementing sustainable bioenergy production: A compila‐tion tools and approaches. IUCN, Gland, Switzerland; 2008.


[55] Geyer R, Lindner JP, Stoms DM, Davis FW, Wittstock B. Coupling GIS and LCA forbiodiversity assessments of land use. International Journal of Life Cycle Assessment2010; 15(7): 692-703. doi: 10.1007/s11367-010-0199-9

[56] Neupane B. Incorporating biodiversity impact into environmental life cycle assess‐ment of woodchips for bioethanol production, FTY2011-001; http://www.library.umaine.edu/theses/theses.asp?Cmd=abstract&ID=FTY2011-001 (ac‐cessed 16 February 2011) University of Maine; 2011

[57] Kloverpris J, Wenzel H, Banse M, Milai Canals L, Reenberg A. Global land use impli‐cations of biofuels: state-of-the-art. Conference and workshop on modeling globalland use implications in the environmental assessment of biofuels. InternationalJournal of Life Cycle Assessment 2008; 13 (3): 178-183

[58] Holmberg J, Robert JH. Backcasting from non-overlapping sustainability principles—A framework for strategic planning. International Journal of Sustainable Develop‐ment and World Ecology 2000; 7:1-18.

[59] Holmberg J, Lundqvist U, Robert K, Wackernagel M. The ecological footprint from asystems perspective of sustainability. International Journal of Sustainable Develop‐ment and World Ecology 1999; 6:17-33.

[60] Kuisma J. Backcasting for Sustainable Strategies in the Energy Sector. IIIEE Report.Lund, Sweden, September 2000; Volume 18.

[61] Robinson J. Future subjunctive: Backcasting as social learning. Futures 2003; 35:839-856.

[62] Geurs K, Wee VB. Backcasting as a tool to develop a sustainable transport scenarioassuming emission reductions of 80–90%. Innovations 2000; 13:47-62.

[63] Dreborg KH. Essence of backcasting. Futures 1996; 28:813-828.

[64] Borbely M, Muirhead JR, Graniero PA. Backcasting and forecasting biological inva‐sions of inland lakes. Ecological Applications 2004; 14:773-783.

[65] Baumgartner, R. Dealing with uncertainty—integrated sustainability assessmentbased on Fuzzy Logic Sustainable Development and Planning II; WIT Press: Billerica,MA, USA; 2006 Volumes 1, 2.

[66] Law A, Kelton W. Simulation Modeling and Analysis, 3rd ed.; McGraw-Hill: NewYork, NY, USA; 2002.

[67] Kempner R, Beck J, Petrie J. Design and analysis of bioenergy networks: a complexadaptive system approach. Journal of Industrial Ecology 2009; 13: 284-305.

[68] Halog A, Sagisaka M, Inaba A. Modeling uncertainties in assessing waste gasificationtechnology. MACRO Review Special Issue 2003; 16:251-255.

[69] Huijbregts M. Dealing with parameter uncertainty and uncertainty due to choices inlife cycle assessment. International Journal of Life Cycle Assessment 1998; 3:343-351.


215

[70] Ford A. Modeling the Environment: An Introduction to System Dynamics Modelingof Environmental Systems; Island Press: Washington, DC, USA 1999; p3

[71] North MJ, Collier NT, Vos JR. Experiences creating three implementations of the re‐past agent modeling toolkit. ACM Transaction on Modeling and Computer Simula‐tion 2006; 16:1-25.

[72] Miller J. Active nonlinear tests (ANTs) of complex simulation models. ManagementScience 1998; 44: 620-630.

[73] Scholl H. Agent-based and system dynamics modeling: A call for cross study andjoint research. http://projects.ischool. washington.edu/jscholl/Papers/DTABS01.PDF(accessed 16 February 2011). In Proceedings of 34th Annual Hawaii InternationalConference on System Sciences, Maui, HA, USA, 3–6 January 2011; 2011.


the need for integrated life cycle sustainability analysis ......chapter 7 the need for integrated...

Documents